Article Text
PDF
Abstract
Objective LV longitudinal strain, a recognised marker of LV function, has been recently applied to the evaluation of the athlete's heart. At present, little is known about the influence of training on LV global longitudinal strain (GLS) in athletes. The aim of this study was to prospectively investigate the impact of training on LV longitudinal strain and twist mechanics in a cohort of competitive athletes.
Methods Ninety-one competitive athletes, practising team sports and competing at national or international level, were analysed. Echocardiographic evaluation was performed at the beginning of the season (low training) and after 18±2 weeks of a supervised, intensive training programme (peak training).
Results A significant increase in LV mass (p<0.0001), LV end-diastolic and end-systolic volume (p=0.0001 and <0.0001, respectively) was found at peak training. LV basal and apical torsion (p=0.59 and 0.43, respectively) and LV twisting (p=0.78) did not change, and only a mild increase in LV GLS was evident after training (p=0.044). Resting heart rate was identified as the only independent predictor of LV GLS after training (β=0.30, p=0.005).
Conclusions A 18-week, intensive training programme induced only a slight increase in LV GLS despite marked changes in cardiac morphology, suggesting a physiological adaptation of the LV to exercise conditioning.
Statistics from Altmetric.com
Introduction
LV adaptations to exercise conditioning have been extensively characterised in the past years and referred as ‘athletes’ heart’.1 ,2 Cross-sectional studies have demonstrated that training-induced LV remodelling encompasses both morphological and functional adaptations and is dynamic in nature, with the most evident changes occurring during the peak season, even in subjects already presenting the features of athlete's heart.3–6
The 2D speckle-tracking echocardiography (STE) has been recently applied to characterise cardiac function in athletes, showing an improvement in LV twist mechanics in athletes with LV remodelling compared with sedentary controls.7 ,8 However, at the present, only a few data are available regarding the time and extent of training-induced changes in myocardial deformation as analysed by STE.9 ,10 In view of the growing application of 2D STE on the integrated imaging-based diagnosis of cardiomyopathies, understanding the influence of exercise training on LV strain is now timely and appropriate. Moreover, while cardiac functional changes have been extensively investigated in endurance athletes, a few data are available in athletes engaged in team sports, that is, the most popular sports in Europe, such as soccer, basketball or volleyball.11
Therefore, the aim of this study was to prospectively assess the impact of a supervised training programme on LV strain and twist mechanics by using 2D STE in professional athletes engaged in team sports. We hypothesised that, despite changes in LV size, LV strain would not change in young, healthy athletes.
Methods
Study population
We enrolled 101 professional athletes engaged in soccer, basketball and volleyball. Athletes, competing in a national or international level, participated in a supervised intensive training programme. Ten athletes were eventually excluded because they were withdrawn from the training programme (for >15 days) due to musculoskeletal injuries; thus, the final study population consisted of 91 athletes.
Echocardiographic examination was performed at the beginning of the training season, before the competitive period (ie, low-training period), while the evaluation at peak training was performed 18±2 weeks after, that is, during the competitive phase.
During the low-training period, athletes trained for about 20 h/week. Training sessions consisted of high-volume/low-intensity running and sprinting conditioning. Athletes also performed 2–3 resistance-training sessions/week at high-moderate workload.
During the peak training period, athletes trained for at least 12 h/week and played one/two matches a week. They trained at workloads from 70% to 95% of maximal heart rate (HR), as indicated by individual HR monitoring. Training sessions consisted of technical-tactical drills, low-volume/high-intensity running and sprinting conditioning (basketball and soccer players) and jumping and sprinting conditioning (volleyball players). Athletes also performed 1–2 resistance-training sessions per week at moderate workload.
All participants underwent complete physical examination, ECG, echocardiography and treadmill ECG test with no evidence of pathological cardiac findings. Systolic and diastolic blood pressure, body weight and height were measured and body surface area (BSA) was calculated.12 All participants were free of manifested coronary artery disease, valvular and congenital heart disease, cardiomyopathy, arterial hypertension and diabetes mellitus. All subjects were asymptomatic and did not report family history for sudden cardiac death.
Standard echocardiographic assessment
Echocardiographic examination was performed by one cardiologist using a high-quality echocardiograph (Vivid 9, GE, Milwaukee, Wisconsin, USA), equipped with an M5S 1.5–4.0 MHz transducer, and a one-lead ECG was continuously displayed. Offline data analysis, from three stored cycles, was performed by an experienced reader, blinded to the study timepoint, using a dedicated software (EchoPac, V.112, GE, USA). HR was measured from the superimposed ECG obtained during the echocardiographic examination.
Chamber's size quantification
Linear measurements of interventricular septum, posterior wall thickness and LV internal diameters were obtained from the parasternal long-axis acoustic window. LV internal diameters and wall thicknesses were measured in parasternal long-axis view, at the level of LV minor axis, as recommended.13 LV mass was calculated from the American Society of Echocardiography recommended formula.13 Relative wall thickness was calculated by the formula: (2×posterior wall thickness)/LV internal diastolic diameter.13 LV volume measurements were calculated from the apical views using the modified Simpson's rule and EF was calculated as (EDV−ESV)/EDV, where EDV is end-diastolic volume and ESV is end-systolic volume.13
Left atrial volumes were calculated by modified Simpson's rule in the apical four-chamber and two-chamber views and an average value was obtained.13
Standard and tissue Doppler imaging
Pulsed-wave Doppler was performed in the apical four-chamber view to obtain mitral inflow velocities, as recommended.14 Measurements of mitral inflow included the peak of early filling (E-wave) and late diastolic filling (A wave), and the E/A ratio.
Tissue Doppler imaging was performed placing the sample volume at the level of septal and lateral insertion sites of the mitral leaflets from the apical four-chamber view,14 obtaining peak systolic (s′), early diastolic (e′) and late diastolic (a′) annular velocities. The average value of septal and lateral velocities was calculated. The e′ velocity and the derived e′/a′ ratio were used as markers of ventricular diastolic relaxation.14 The E/e′ ratio was also obtained and used as an index of intra-cardiac filling pressures.15
Speckle-tracking echocardiography
The 2D STE analysis was performed on narrow-sector greyscale images of LV from apical and parasternal short-axis views with temporal resolution of 60–90 frames/s. All images were optimised with gain, compression and dynamic range to enhance myocardial definition with standardised depth, frequency and insonation angle for all participants.16 Offline analysis was performed by one experienced reader using a commercially available semi-automated 2D strain software (EchoPAC PC, V.112, GE, USA).
After manual tracing of endocardial border, the software automatically provided myocardial motion and divided the LV into six regions of interest, accepting segments of good quality tracking while rejecting poorly tracked segments, and allowing the observer to manually override its decision by using visual assessment. Results were averaged on three cardiac cycles. Global longitudinal LV strain (GLS) was assessed from six segments in the posterior and anteroseptal walls (apical long-axis view), posterior septal and lateral walls (four-chamber view), and inferior and anterior walls (two-chamber view). GLS was calculated by averaging all values of regional peak longitudinal strain, obtained in each view before aortic valve closure. Basal-to-apical torsion was calculated as the net difference in LV rotation angle at the apical (counterclockwise) and basal (clockwise) short-axis plane occurring at end-systole. By convention, counterclockwise rotation, as viewed from LV apex, was marked as a positive value, whereas clockwise rotation was expressed as a negative value. Rotation angles were expressed in degrees. The basal plane was defined as the highest basal imaging plane at which uniform full thickness myocardium was observed surrounding the mitral valve at end-systole. The apical level was chosen as the imaging plane with no visible papillary muscles that most closely approximated an end-diastolic ratio of LV cavity diameter to total LV diameter of 0.5.9
LV twisting curve was calculated as the net difference between peak systolic apical and basal rotation. LV torsion was obtained indexing LV twisting to LV length, measured in apical four-chamber view and defined as the end-diastolic length from the mitral valve hinge point plane to the most distal endocardium at the LV apex.9 The degree of untwisting rate was defined as the directional reversal of systolic counterclockwise twist during diastole and expressed as percentage of untwisting.17
Statistical analysis
The sample size was calculated using a significance level of 0.05 and a power of 0.80 to detect an assumed mean change of 2%±3% in global LV strain on the basis of literature data. A group of >85 competitive athletes were deemed adequate to meet the aforementioned criteria. Normal distribution of all continuous variables was examined using the Shapiro–Wilk test. The paired t test and the Wilcoxon matched-pair test were used to assess the within-subjects significance of pre-training and post-training measurements, as appropriate for data distribution (as reported in the tables). A p value <0.05 was considered significant. Univariable and multivariable linear regression analyses were used to examine independent predictors of LV GLS, LV mass and LV untwisting rate. LV mass, LV EDV and ESV, cardiac output, LV GLS, E/A ratio, E/e′ ratio, e′/a′ ratio, resting HR, systolic and diastolic blood pressure were entered in the univariable analysis. The multivariable model included variables identified as statistically significant at the univariable analysis. Statistics were performed using SPSS V. 22.0 (Statistical Package for the Social Sciences, Chicago, Illinois, USA).
Results
Demographic characteristics of the study population are reported in table 1.
Characteristics of the study population
The mean age of the study population was 23±6. Thirty-six (40%) of the athletes were female. BSA showed a modest, significant increase after training, driven by an increase in body weight (p=0.004 vs pre-training). Resting HR decreased compared with baseline data (p<0.0001).
LV remodelling
LV morphology changes observed at peak training are reported in table 2 and figure 1.
Comparison between low-training and peak training LV parameters
Changes in LV dimensional parameters observed at low-training and peak training timepoints in competitive athletes. See text and table 2 for details.
Both LV end-diastolic dimension and wall thickness significantly increased at peak training, resulting in an increase in LV mass compared with low-training period (p<0.0001). The relative LV wall thickness increased after training (p=0.011). Both LV end-diastolic and end-systolic volumes increased (p=0.001 and <0.0001, respectively), while LVEF did not vary (p=0.68). Moreover, a significant increase in left atrial volume was also found at peak training (p<0.0001).
The increase in LV chamber's size was associated by an improvement in LV diastolic function, as suggested by the increase in E/A and e′/a′ ratios (p=0.034 and <0.0001, respectively). Conversely, no changes were observed in LV filling pressures (estimated at rest by E/e′ ratio) between low and peak training (p=0.89).
LV strain and twist mechanics
A slight improvement in LV GLS was found in athletes after training (p=0.044), as reported in table 3 and figure 2A. The maximum values of GLS observed at low training and at peak training timepoint were −25.8% and −27.2%, respectively, while the minimum values were −15.2% and −15.1%, at low-training and peak training timepoint, respectively.
Comparison between low training and peak training in LV 2D speckle-tracking-derived parameters
Changes in LV longitudinal strains observed at low-training and peak training timepoints in competitive athletes. (A) LV global longitudinal strain; (B) LV strain for basal, mid and apical segments, respectively. See text and table 3 for details.
LV longitudinal strains tended to increase after training with a significant difference observed for basal and mid-cavity longitudinal strains (p=0.034 and 0.037, respectively) (figure 2B). LV twist mechanics were analysed in all subjects. Neither basal nor apical rotation did change at peak training (p=0.59 and 0.43, respectively). Similar results were found also for LV torsion, twisting and untwisting rate (p=0.42, 0.78 and 0.55, respectively).
Predictors of LV strain and LV mass
Resting HR (β=0.30, p=0.005), LV mass (β=−0.24, p<0.05) and LV EDV (β=−0.26, p<0.05) were identified as predictors of LV GLS at peak training, with resting HR being the only independent predictor at multivariable regression analysis (table 4).
Predictors of LV functional and dimensional parameters
Indeed, resting HR (β=0.35, p=0.001), E/e′ ratio (β=0.25, p<0.05) and cardiac output (β=0.23, p<0.05) were identified as predictors of LV untwisting rate with resting HR being the only independent predictor at multivariable regression analysis (table 4).
In addition, BSA (β=0.73, p<0.0001), resting HR (β=−0.29, p=0.006), LV twisting (β=0.29, p=0.008) and cardiac output (β=0.61, p<0.0001) were identified as independent predictors of LV mass at peak training, with cardiac output being the only independent predictor at multivariable regression analysis (table 4).
Discussion
Implementation in the clinical practice of 2D STE, and particularly of LV GLS, is rapidly growing due to the ability of this novel echocardiographic technique to characterise LV hypertrophy.18 ,19 For instance, in genotype-positive, phenotype-negative hypertrophic cardiomyopathy (HCM) individuals, the 2D STE is capable of detecting subtle functional anomalies even before the onset of LV hypertrophy.20 Furthermore, when pathological hypertrophy is present, myocardial longitudinal velocities and deformation parameters may be reduced, with longitudinal deformation being typically diminished at the site of hypertrophy, despite preserved global LV function as commonly assessed by EF.21 This technique has, therefore, the potential to convey useful clinical information, and it begins to be deemed in the differential diagnosis of HCM from physiological LV remodelling in athletes.
We planned the present investigation to ascertain whether, and to what extent, the LV strain is affected by intensive exercise training during the competitive season in athletes. Our study demonstrates that a short period (18 weeks) of intensive athletic conditioning induces only slight changes in LV GLS despite a significant morphological remodelling. Only a few longitudinal studies are available in athletes, and, to our knowledge, this is the first longitudinal study examining experienced athletes during a period of intensive training.
Our observation that exercise training induces only slight changes in LV GLS, a functional parameter expression of the subendocardial fibres mechanics,16 is in agreement with observations in animal studies demonstrating that the effect of exercise training in contractile properties is more pronounced in the subendocardial than subepicardial region.22 ,23 In humans, Leggio and colleagues observed in hypertensive patients that LV GLS significantly increased after training,24 while Hansen et al25 found that a 3-month training programme was not able to alter LV GLS in overweight preadolescent children engaged in a competitive football season. Conversely, a previous study by Weiner et al,9 investigating the impact of a 90-day period of training on LV twist mechanics in university male rowers, showed an increase in LV apical rotation, twisting and untwisting rate after training. A possible explanation for these discrepancies relies on differences in demographic and training characteristics of the study populations: athletes enrolled by Weiner were endurance, younger and non-professional university level, contrary to our population of competitive athletes who have undergone high-volume training programmes for many years. Taking together, these differences suggest that, while LV twist mechanics can change with training in young, not-experienced athletes, it does not vary with training in a population of experienced athletes where changes would be expected to be minimal with continued training.
A relevant consideration that rises from our study is that LV myocardial deformation dynamics were not impaired by intensive exercise training, supporting the hypothesis of a physiological adaptation of the LV to training. Notably, GLS was within the normal range in athletes, suggesting that a reduction in LV GLS is an uncommon feature in athlete's heart.
It is known that a variety of determinants might potentially influence LV strain, including demographic features (gender, ethnicity, anthropometric variables), haemodynamic factors (heart rate, blood pressure) and cardiac morphology (LV size and wall thickness). In this study, we found that resting HR, LV mass and LV EDV were determinants of LV GLS, with resting HR being the only independent predictor. A relationship between LV GLS and resting HR was not found in previous studies analysing the general population 26 and not confirmed in the athletes.7 ,8 ,27 These findings can be explained, at least in part, by the fact that athletes with low-resting HR show also great cardiac dimensions, suggesting that not only resting HR but also cardiac size influence LV longitudinal strain. We observed a decrease in longitudinal strain values with increasing cardiac dimensions, and these findings are in agreement with experimental data by Rösner and colleagues, who demonstrated a decrease in LV longitudinal strain with increasing heart size, with a hyperbolic relationship independent by heart shape and global contractility.27
The participation in different sports disciplines could probably affect strain values, also depending on the specific changes induced on LV geometry. This is probably the reason why we observed higher LV GLS values compared with previous data by D'Andrea et al28 and Caselli et al,29 who studied different athletic populations. Although all these potential determinants of LV strain should be taken into account and further studies investigating different sports disciplines are needed, a GLS <−14% or <−15% (depending on the brand of imaging equipment) is an uncommon feature of athlete's heart.7 ,8 ,29
In a large population of healthy individuals, gender and age were identified as predictors of LV strain, after adjusting for blood pressure, LV cavity dimension, body mass index and non-fasting serum glucose.30 Conversely, in the present study, in a population of healthy subjects, despite a significant increase in body weight and relative wall thickness after training, we found a mild increase in LV GLS rather than a decrease. Furthermore, no significant differences in LV strain were observed between female and male athletes (data not shown). Our data are in agreement with a recent meta-analysis reporting that the effects of age, gender and body mass index had no significant impact on changes of LV GLS.26 Thus, although the identification of the factors influencing LV strain was beyond the scope of the study, our findings confirm that, in healthy individuals engaged in team sports, LV strain does not significantly differ between male and female, neither before nor after training.
Study limitations
The lack of data quantifying exercise capacity (VO2 max) before and after the training programme prevents us from assessing a direct relationship between changes in exercise capacity and LV deformation mechanics. This represents an important area of our ongoing work. Second, only athletes practising team sports were enrolled in this study and athletes engaged in disciplines with high cardiovascular impact, such as endurance sports, were not included in the analysis; therefore, our results may not be transferable to the entire population of competitive athletes.
Conclusions
Participation in a 18-week intensive training programme was associated only with a slight increase in LV GLS in competitive athletes engaged in team sports, despite a significant increase in LV cavity size. Thus, training is associated with a significant morphological remodelling of the LV, but does not cause a reduction in LV GLS, suggesting that a decreased LV GLS is an uncommon feature in athlete's heart and cannot be considered as a physiological adaptation to training, irrespective of the in-seasonal period when the athlete is evaluated.
Key messages
What is already known on this subject?
Speckle-tracking echocardiography has been previously applied to the ‘athlete's heart’ in cross-sectional studies, demonstrating a physiological adaptation of the LV to exercise conditioning. However, little is known about the possible influence of training during the training season. Our hypothesis was that, despite changes in LV size, LV strain would not change in physiological cardiac remodelling.
What might this study add?
In this longitudinal study, we demonstrated that, despite a training-induced increase in LV dimensions, LV longitudinal strain did not decrease. Conversely, after an 18-week, intensive training programme, a slight increase was found in LV longitudinal strain (p=0.044).
How might this impact on clinical practice?
Training does not affect LV strains in competitive athletes, neither at the beginning nor during the season. These findings suggest that physiological adaptation of the LV to exercise conditioning is not associated with a reduction in LV strain.
References
Footnotes
Contributors FDA and AP wrote the manuscript. FDA, MB and SM contributed to the conception and to the design of the study. FDA, FA, MS, AL, MC and MF performed data acquisition, analysis and interpretation of data. AP, MB and SM critically revised the manuscript.
Competing interests None declared.
Patient consent Obtained.
Ethics approval The protocol was planned in accordance with the ethical standards of the 1964 Declaration of Helsinki and later amendments and study was approved from the Local Ethical Committee.
Provenance and peer review Not commissioned; externally peer reviewed.